Strange SCX retention

[Chemist Waldron]

Good morning,
I am running the EP method for copper tetraMIBI related substances. Copper tetraMIBI is a Cu complex used in preparation of a pharmaceutical. The LC method is SCX using an Agilent Zorbax.
20µL of the test solution (50 mg/mL) is injected and any impurities found are quantified using the API peak from a 20µL injection of a ref standard, which is a 1000x dilution (in MP) of the test solution. The ref solution is injected first.

(see attached jpg) - Peak #2 is the normal 20µL inj of the ref solution. Peak# 4 is the 20µL inj of the test solution. #4 peak shape is fine (albeit topping out) and elutes at the expected time of about 5 mins. #2 has a shark's fin front and elutes early. #6 is a 100µL inj of the ref soln, it shows slightly more retention. #7 is a 0.1µL injection of the test solution. These chromatograms show that the peak shape and retention deteriorate when lower quantities of API are injected, irrespective of solution concentration.

Anyone have any idea what is going on here?





Thanks

[tkubowicz]

Hello

Even if you overload column and peak shape is "triangle" you should have consistent retention time. So if all chromatograms are in the same time scale there is something wrong.
I'd check:
1. Pump - problems with valves/gradient valve (wrong gradient formation)
2.Mobile phase - check pH (prepare new mobile phase)

Regards

Tomasz Kubowicz

[mattmullaney]

Hello,

I agree with Thomas--I'd suspect the mobile phase first, this peak shape could be the result of some sort of chemical equilibrium problem with the Copper complex as it passes through the column...this may also be why retention may be affected.

As to this method in the EP, is the compound name Copper tetramibi tetrafluoroborate? I'm having a time finding the method in the EP altogether.

[mattmullaney]

Apologies...for example, the monograph number for Copper tetramibi tetrafluoroborate would help me in my search. Can't get the hang of looking in the EP for some reason.

[Chemist Waldron]

Here is the analyte and method. (Edit: this is a complex; the structure is incorrectly shown with covalent bonds to the Cu). Sorry, it was too small to read in the OP.
So the strange thing is that the fronting and early elution occurs when the 1000x dilution is injected. I have checked the MP prep and repeated the test months later. Thanks.



Related substances. Liquid chromatography (2.2.29). Prepare the solutions immediately before use.
Test solution. Dissolve 50.0 mg of the substance to be examined in the mobile phase and dilute to 10.0 mL with the mobile phase.
Reference solution. Dilute 1.0 mL of the test solution to 100.0 mL with the mobile phase. Dilute 1.0 mL of this solution to 10.0 mL with
the mobile phase.
Column:
– size: l = 0.25 m, Ø = 4.6 mm;
– stationary phase: strong cation-exchange silica gel for chromatography R (5 μm);
– temperature: 20 °C.
Mobile phase: mix 40 volumes of acetonitrile R1 and 60 volumes of a 20.4 g/L solution of potassium dihydrogen phosphate R,
previously adjusted to pH 6.0 with a 56.1 g/L solution of potassium hydroxide R.
Flow rate: 1.0 mL/min.
Detection: spectrophotometer at 230 nm.
Injection: 20 μL.
2547E.htm
2547E.htm[3/1/2017 2:36:20 PM]
Run time: twice the retention time of copper tetramibi.
Retention time: copper tetramibi = about 5 min.
System suitability: reference solution:
– symmetry factor: maximum 1.35 for the principal peak.
Calculation of percentage contents:
– for each impurity, use the concentration of copper tetramibi in the reference solution.
Limits:
– unspecified impurities: for each impurity, maximum 0.10 per cent;
– total: maximum 0.2 per cent;
– reporting threshold: 0.05 per cent.

[mattmullaney]

Hi Chemist Waldron,

Interesting. So the behavior you're seeing happened more than once as well? Looks like a chemical problem to me, still. Are you using a Needle Wash for this method, and if so, what is its composition? Perhaps residual needle wash solvent is "throwing off" the complex when it is diluted in such a way that is not manifested when the large concentration injection is made.?

[Chemist Waldron]

The rinse is 10% ACN in water. This was the rinse the previous time, same LC. Since the MP is 40% ACN, would the theory be that water is the interferent in the rinse and 40% ACN in water would be better? The system is a Shimadzu Prominence.

[mattmullaney]

Hi Chemist Waldron,

My supposition is that is the case, sort of. For this separation, I think that the Needle Wash of 10:90 ACN/water may be sufficiently different from the mobile phase that, if it remains in the injector, may contribute to equilibrium trouble for the analyte peak when it is dilute. Both pH and aqueous content are different...please also, what is the flow rate and the column dimensions?

[mattmullaney]

Oh geez, my apologies. Column dimensions are 4.6 x 250 and flow rate is 1 mL/min, so the column void time may be estimated as ca. 2.8 minutes...a retention factor of 3 would be roughly 8.5 minutes. These peaks, diluted or not, are eluting pretty early in the separation.

[tkubowicz]

Hello

Have you checked pump? I still think you've got problem with it.
Are you running mobile phase pre-mixed or you mixing ACN and buffer online? Check your pump for leaks (run diagnostic to see if valves are working fine).

Regards

Tomasz Kubowicz

[Chemist Waldron]

Tomasz - The pump is working ok and I have eliminated this as a route cause because the retention drift is correlated to the amount of analyte injected. The method is shown above along with my initial troubleshooting comments.

[mattmullaney]

Good Afternoon,

Hmm. This reminds me a little bit of the method for nitrilotriacetic acid in EDTA that I tried out back at Merck a while ago. That was a RP method as opposed to a SCX one, my recollection is fuzzy but that Cu(II)-NTA complex had a fronting shape as I recall, similar to this analysis...and also as a low level. Wish I had a better memory of the details. Also as the separation was/is RP, it may not be a perfect analogy of what happens in this method.